Help:Normalize EoM references

Perl script to normalize old EoM reference tables

The perl script below can be used to transcribe old <table> based reference lists into wikipedia table style, thereby using reference descriptors based on the authors' names, ordering tables alphabetically by authors' names, and installing links and anchors using Template:Ref and Template:Cite.

If an external editor is used to edit EoM pages, then its application should be easy. For example, if the external editor is emacs, and if the perl script code below is stored in some file

~/encyclopedia/references/eom_refs.pl

then, after loading the page code into emacs (for example by using the "It's All Text" add-on), the following code (below named Emacs macro) inserted in the user's .emacs file will reduce the application of the perl script to just one keystroke "Ctrl ," (that is, simultaneously press the control key and the comma key).

Emacs macro

;; Requires filter eom_refs.pl to rewrite table references into
;; wikipedia style tables
;; 

(defun eom-references () 
  (interactive)
  (let ( (coding-system-for-read 'utf-8) 
	 (coding-system-for-write 'utf-8)
	 )
    (shell-command-on-region 
       (point-min) (point-max) 
       "~/encyclopedia/references/eom_refs.pl" nil t)
    )
  )

(global-set-key (kbd "C-,") 'eom-references)

Perl script

#!/usr/bin/perl -w

# This script operates like a filter, i.e, it reads a file from
# standard input, rewrites its content and writes the transcript to
# standard output.

# It rewrites eom reference tables into wikipedia table style. It uses
# the Ref and Cite templates.

# It replaces the reference descriptors (like [1], [2],
# etc. by the first two letters of the author names like [Ab], [Ha],
# [KuZi] etc.  and orders the tables alphabetically with respect to
# these descriptors.  Moreover, in the tables, anchors are set by
# {{Ref|Ab}} using the Ref template while, within the text, references
# {{Cite|Ab}} to these anchors are installed by the Cite template.

# Author names are retrieved by the following heuristic: 

# All those words after the ref descriptor (in brackets) are collected
# which start with a capital followed by a small letter and ended by a
# comma (like Abel, ), and the first two letters of all these words
# are concatenated and used as new reference descriptor [Ab]. 

# In case of multiplicities of such strings the second, third, ... gets a
# '2','3', ... appended like [Ab2], [Ab3] etc.


use strict;
use utf8;

# get file content:
undef $/;
my $f = <>;

# collect all reference lists in array
my @refs = ($f =~ /\=+References\=+\s+(\<table\>.*?\<\/table\>)/sg);
if ($#refs < 0) {
    print "Error:  no refs: length of refs array: $#refs\n"; exit(1);
}

my %H = (); # Hash for ref keys like [1] etc.
my %K = (); # Hash for multiplicity of ref key occurence

# Array for all initial keys:
my @items = ($f =~ /\>\[(\w+?)\]\<.*?,/sg);
foreach(@items) { 
    my $key = $_;
    my $y = "";
    # find names after $key and before next comma:
    if ($f =~ /\>\[$key\]\<(.*?),\s*["\']/sg) {
	my $x = $1;
	# collect first two letters of names 
	# and concat into $y, will serve as new key:
	my @A = ($x =~ / ([A-Z]\w)[^\.]/sg);
	foreach(@A) { $y .= $_;	}
    }
    $K{$y}++; 
    if ($K{$y} > 1) { # append count if bigger than 1
	$H{$key} = $y.$K{$y};
    } else {
	$H{$key} = $y;
    }
}

$f =~ s/\[\[#References\|\[(\w+?)\]\]\]/{{Cite|$H{$1}}}/sg;
$f =~ s/\|\s*\[(\w?)\]\s*\|/|{{Ref|$1}}|/sg;


foreach(@refs) { # process all ref lists:
#    print "$_\n";
    my $x = $_;
    # do table conversion:
    my $y = &replace($x);
   foreach(keys %H) {
	$y =~ s/\[$_\]/{{Ref|$H{$_}}}/sg;
    }
    # sort bib entries by ref keys 
    my %R = ();
    my @A = split(/\|\-/,$y);
    foreach(@A) {
	if (/\{\{Ref\|(\w+)\}\}/) { $R{$1} = $_; }
    }
    my $z = '{|'."\n".'|-';
    foreach(sort keys %R) { $z .= $R{$_}.'|-';  }
    $z .= "\n\|\}\n";

    # do replacements:
    # prepend a backslash before each of <>"=()[]+-?|^$*\~
    $x =~ s/([\<\>\"\=\(\)\[\]\+\-\?\|\^\$\*\\\~])/\\$1/sg;
    $f =~ s/$x/$z/sg;
}

#print "...............\n";
print $f;

exit(0);

## converts bib table from html to wiki structure
sub replace{ 

## cf. http://en.wikipedia.org/wiki/Help:Table#Pipe_syntax_tutorial
    my $f = $_[0];
    # replace data cells (last | still to be removed!):
    $f =~ s/\<td (.*?)\>(.*?)\<\/td\>/|$1|$2|/sgi;
    # replace row (consider removal of last | from above):
    $f =~ s/\<tr\>(.*?)[\s\|]*\<\/tr\>\s*/|\-\n$1\n/sgi;
    # no spaces between consecutive |:
    $f =~s/\|\s*\|/||/sgi;
    # replace <table> and </table>:
    $f =~ s/\<table\>/\{|\n/sgi;
    $f =~ s/\<\/table\>/|\-\n|\}/sgi;
    # remove possible | before template calls:
    $f =~ s/\|\s*\{\{/\{\{/sg;
    # remove space before some punctuation:
    $f =~ s/ +([\,\.])/$1/sg;
    return $f;
}